Remei Ridorsa – Gustavo Patow Geometry and Graphics Group Universitat de Girona T HE SKYLINE E...

Preview:

Citation preview

Remei Ridorsa – Gustavo PatowGeometry and Graphics GroupUniversitat de Girona

THE SKYLINEENGINE SYSTEM

http://iiia.udg.edu/GGG/skylineEngine/

INDEX

Introduction Previous work Pipeline Main modules Other modules Implementation Results Future work

2

INTRODUCTION

Procedural modelling: creating models from rules instead of geometric primitives

More efficient than traditional modelling Procedural modelling of cities:

Is necessary for: Architectonic and urban design Films and games Virtual reality

Is an open problem

3

AIM

Procedural tool for city generationIn an easy way User friendlyUsing rules defining the city

4

FEATURES

Features available on high-end commercial systems: Street pattern based tools to design urban

layouts(grid, organic, radial)

Real street networks of any city can be imported Images OpenStreetMap

buildingEngine, building definition module Other modules Parameters controlled globally via image maps

5

CONTRIBUTIONS

Contrary to the accepted standard linear pipeline, we have developed a non-linear one

We have created a configurable and flexible node-based modelling system

We have developed a new method to subdivide blocks into lots in the style of European cities like Paris or Barcelona. 6

PREVIOUS WORK

[Parish 01] First procedural approach, based on L-systems

[Sun 02] Template-based street network geneation

[Chen 08] Tensor fields for street generation

[Wonka 03] [Müller 06] [Müller 07] [Lipp 08] Procedural methods for building and facade generation using rule-based rewriting systems

7

TRADITIONAL PIPELINE The traditional pipeline is not enough for a

complete complex city.

As an example, we can have parks at different levels in the pipeline.

This results in an Acyclic Directed Graph (DAG). 8

SKYLINEENGINE PIPELINE

9

MAIN MODULES: INPUT

In an XML file: For synthetic cities:

District map Avenues pattern Land use map Heights map

For real cities: City map (image) OpenStreetMap

10

MAIN MODULES: DISTRICTS

Districts with different patterns

11

currMap = Øfor each pattern є districtMap do

Generate patternPosition and orient patternbounds = Bounds(pattern, districtMap)finalP = Clip(pattern, bounds)finalP = Merge(finalP, bounds)currMap = Merge(finalP, currMap)

end forreturn currMap

MAIN MODULES: MAJOR STREETS

12

Grid Pattern

Eixample Barcelona

Radial Pattern

Karlsruhe (Germany)

MAIN MODULES: MAJOR STREETS

13

Organic Pattern (voronoise)• We use a Voronoi variant (Lloyd’s algorithm)• Density can come either from a map or from an

analytic function

MAIN MODULES: MAJOR STREETS

Image processingEdge selectionBinarization

If we have a district map: We do the same processing generate streets

between districts

14

MAIN MODULES: MAJOR STREETS

If we have a land-use mapShow were we cannot build buildings

15

MAIN MODULES: MINOR STREETS

Number of minor roads proportional to the block area

Create a grid with this number of divisions Rotate grid Do an intersection between both

16

MAIN MODULES: BLOCKS

For OSM and streets based on paths, the resulting network is a graph

Blocks are minimal enclosing circuits

17

MAIN MODULES: LOTS

Manhattan-style: Split block longitudinally Crete a given number of lots in each partition

18

MAIN MODULES: LOTS

Paris-style:Round anglesSubdivide into a number of lots

proportional to the block perimeterCreate inner primitive

19

MAIN MODULES: BUILDINGS

20

Height map

21

OTHER MODULES: BUILDINGENGINE

OTHER MODULES: 4D CITIES

22"Interactive Geometric Simulation of 4D Cities", Basil Weber, Pascal Mueller, Peter Wonka and Markus Gross, Eurographics 2009

OTHER MODULES:EXAMPLE-BASED URBAN LAYOUT SYNTHESIS

23“Interactive Example-Based Urban Layout Synthesis”, D. Aliaga, C. Vanegas i B. Benes, SIGGRAPH ASIA 2008

0.05

0.280.24

0.07

0.00

0.000.00

OTHER MODULES: STREET CONTENT

24

OTHER MODULES: BRIDGEE

25

OTHER MODULES: TRAFFICSIM

26

IMPLEMENTATION: HOUDINI 3D modeler Procedural behaviour Uses a hierarchical node system Programmable in Python

27

RESULTS: PIPELINE

28

29

RESULTS: MANHATTAN-LIKE

RESULTS: GIRONA

30

Synthetic buildings

FUTURE WORK

Exact localization and manipulation of landmarks, etc.

Manage level-of-detail Batch exportation Others...

31

ACKNOWLEDGEMENTS Marc Argence Albert Roca Joel Coll Costa Maria del Carme Hijano Domènech Zhi Chao Yin

32

THANK YOU!

“If you ignore the rules people will, half the time, quietly rewrite them”Terry Pratchett, Equal Rites

Recommended